EXPERIENCE OF USING THE NEURAL-NETWORK APPROACH FOR IDENTIFICATION OF MSF DESALINATION PLANTS

Mathematical models of physical and engineering systems are frequently of high dimension, possessing interacting dynamic phenomena. Information processing requirements for experimenting with these models for control purposes are excessive. It is therefore natural to techniques that reduce the computational effort. In this paper, the neural network approach for nonlinear identification of the multistage flash desalination plant (large scale system) is presented through the manipulation of collected input-output data from the plant. Results of utilizing such an approach are also discussed.